Achieving near-zero added latency for any point in time OS kernel-based application replication
One example method includes intercepting an IO issued by an application, writing the IO and IO metadata to a splitter journal in NVM, forwarding the IO to storage, and asynchronous with operations occurring along an IO path between the application and storage, evacuating the splitter journal by sending the IO and IO metadata from the splitter journal to a replication site. In this example, sending the IO and IO metadata from the journal to the replication site does not increase a latency associated with the operations on the IO path.
1. A method, comprising:
intercepting, by an interceptor within an Operating System (OS) kernel, an Input/Output (IO) issued by an application;
writing, by a journal manager within the OS kernel, the IO and IO metadata to a splitter journal in Non-Volatile Memory (NVM);
forwarding the IO to storage; and
asynchronous with operations occurring along an IO path between the application and storage, evacuating the splitter journal by sending the IO and IO metadata from the splitter journal to a replication site.
2. The method as recited in claim 1 , wherein sending the IO and IO metadata from the splitter journal to the replication site does not increase a latency associated with the operations between the application and storage.
3. The method as recited in claim 1 , wherein a splitter within an OS kernel of a Virtual Machine (VM) performs the operation of: forwarding the IO to storage.
4. The method as recited in claim 1 , wherein the IO and IO metadata are transmitted from the splitter journal to the replication site on a path other than the IO path.
5. The method as recited in claim 1 , wherein the IO path comprises a path between the application and a splitter, and a path between the splitter and the storage.
6. The method as recited in claim 1 , wherein the IO and IO metadata are sent from the splitter journal to a replication site as part of a batch of IOs and IO metadata.
7. The method as recited in claim 1 , wherein the NVM is either memory-accessed NVM, or storage-access NVM.
8. The method as recited in claim 1 , wherein the method is performed by a splitter within an OS kernel of a VM, or an OS kernel of a physical machine.
9. The method as recited in claim 1 , further comprising:
receiving first acknowledgement from the storage after the IO has been forwarded to the storage; and
sending a second acknowledgement to the application after the first acknowledgement has been received from the storage.
10. The method as recited in claim 1 , further comprising experiencing a crash of a machine in connection with which the method is performed and, after restart of the machine, resuming evacuation of the splitter journal at a point where evacuation had previously ceased due to the crash of the machine.
11. A non-transitory storage medium having stored therein instructions that are executable by one or more hardware processors to perform operations comprising: intercepting, by an interceptor within an Operating System (OS) kernel, an Input/Output (IO) issued by an application;
writing, by a journal manager within the OS kernel, the IO and IO metadata to a splitter journal in Non-Volatile Memory (NVM);
forwarding the IO to storage; and
asynchronous with operations occurring along an IO path between the application and storage, evacuating the splitter journal by sending the IO and IO metadata from the splitter journal to a replication site.
12. The non-transitory storage medium as recited in claim 11 , wherein sending the IO and IO metadata from the splitter journal to the replication site does not increase a latency associated with the operations between the application and storage.
13. The non-transitory storage medium as recited in claim 11 , wherein a splitter within an OS kernel of a Virtual Machine (VM) performs the operation of: forwarding the IO to storage.
14. The non-transitory storage medium as recited in claim 11 , wherein the IO and IO metadata are transmitted from the splitter journal to the replication site on a path other than the IO path.
15. The non-transitory storage medium as recited in claim 11 , wherein the IO path comprises a path between the application and a splitter, and a path between the splitter and the storage.
16. The non-transitory storage medium as recited in claim 11 , wherein the IO and IO metadata are sent from the splitter journal to a replication site as part of a batch of IOs and IO metadata.
17. The non-transitory storage medium as recited in claim 11 , wherein the NVM is either memory-accessed NVM, or storage-access NVM.
18. The non-transitory storage medium as recited in claim 11 , wherein the operations are performed by a splitter within an OS kernel of either a physical machine or a VM.
19. The non-transitory storage medium as recited in claim 11 , wherein the operations further comprise:
receiving first acknowledgement from the storage after the IO has been forwarded to the storage; and
sending a second acknowledgement to the application after the first acknowledgement has been received from the storage.
20. The non-transitory storage medium as recited in claim 11 , wherein the operations further comprise experiencing a crash of a machine in connection with which the non-transitory storage medium is performed and, after restart of the machine, resuming evacuation of the splitter journal at a point where evacuation had previously ceased due to the crash of the machine.